This brief introduction presents the motivation to create the Q4S protocol to measure e2e quality on real-time.
The problem of network transmission on the Internet( latency, jitter and packet burst, bandwidth and packet loss) affects strongly “new age” services like interactive video applications, virtualized games, or low latency e-auction. The user experience of these new services is damaged by network quality problems. In addition, the transmission problems evolve during session time, and therefore a quality measurement mechanism is needed to know the quality status of the transmission and apply reactively different palliative mechanisms. This is the role of Q4S: real-time measurement of the end-to-end quality.
The approaches to face network transmission problems, are divided into three groups, which are not incompatible among them:
- Deployments on the last mile: certain services such as cloud gaming may deploy pools of servers at last mile, avoiding the use of the Internet. The cost of the deployment is huge, and therefore this strategy compromises the business, raising the price of the delivered service.
- Application level Adaptive mechanisms: like “adaptive streaming” in which bitrate is reduced when packet loss is detected. This solution may evolve to “adaptive encoding” for interactive video services, in which codec could run faster (doing less compression effort) when network latency increases, or compress more strongly when packet loss is detected. These adaptive mechanisms include advanced techniques such as “lag compensation” through predictive execution and “partitioning” strategies for multiplayer games
- Network level Adaptive mechanisms: asking for more quality (bandwidth or priority) to network provider when transmission problems are detected and cannot be solved successfully by application level adaptive mechanisms. This approach could open a new possibility for network operators: to sell dynamic quality on demand to service providers.
For any adaptive mechanisms, it is desirable to have a standard measurement protocol, independent of the type of protocol used in the application. A protocol applicable to measure the end-to-end quality, agnostic to the type of transport protocol used ( HTTP, FTP, RTP, any proprietary protocol, etc.).
The main original contribution of this protocol is to measure the e2e quality connection in parallel to other transport protocols (such as HTTP, RTP, FTP, etc.) without any interferences between transport protocol and measurement protocol.
This measurement protocol is Q4S.